Ebe , Ca a; Klasen, S ephan; Vollme , Sebas ian
A icle — Published Ve sion
Coun ing missing women: a econcilia ion o low and
s ock measu es
Jou nal o Popula ion Economics
Sugges ed Ci a ion: Ebe , Ca a; Klasen, S ephan; Vollme , Sebas ian (2025) : Coun ing missing
women: a econcilia ion o low and s ock measu es, Jou nal o Popula ion Economics, ISSN
1432-1475, Sp inge Be lin Heidelbe g, Be lin/Heidelbe g, Vol. 38, Iss. 4,
h ps://doi.o g/10.1007/s00148-025-01132-0
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/333223
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h p://c ea i ecommons.o g/licenses/by/4.0/
Vol.:(0123456789)
Jou nal o Popula ion Economics (2025) 38:82
h ps://doi.o g/10.1007/s00148-025-01132-0
ORIGINAL PAPER
Coun ing missing women: a econcilia ion o low ands ock
measu es
Ca aEbe 1 · S ephanKlasen2· Sebas ianVollme 2
Recei ed: 2 Augus 2023 / Accep ed: 22 Sep embe 2025
© The Au ho (s) 2025
Abs ac
S ock es ima es o missing women sugges ha excess emale dea hs a e concen-
a ed in Sou h and Eas Asia and among young child en. In con as , low es ima es
sugges ha gende bias in mo ali y is much la ge han p e iously es ima ed using
s ock measu es, is as se e e among adul s as i is among child en in India and China,
and is la ge in Sub-Saha an A ica han in India and China. We show ha he di -
e en s ock and low measu e esul s ely on he choice o he e e ence s anda d o
mo ali y and an incomple e co ec ion o di e en disease en i onmen s in he low
measu e. Al e na i e e e ence s anda ds econcile he esul s o he wo measu es.
Keywo ds Missing women· Gende bias· Mo ali y· Disease· Age· Sub-Saha an
A ica· China· India
JEL Classi ica ion J16· D63· I14· I15· O15
Responsible edi o :Kompal Sinha
S ephan Klasen passed away on Oc obe 27, 2020, a e ba ling he incu able disease Amyo ophic
La e al Scle osis (ALS) o 5 yea s. We a e g a e ul o wha we ha e lea ned om S ephan h ough
ou join wo k and p e iously as s uden s.
* Ca a Ebe
ca a.ebe [email p o ec ed]
Sebas ian Vollme
s [email p o ec ed]
1 RWI – Leibniz Ins i u e o Economic Resea ch, Be lin O ice, Zinnowi ze S . 1, 10115Be lin,
Ge many
2 Uni e si y o Goe ingen, Waldweg 26, Gö ingen37073, Ge many
C.Ebe e al.
82 Page 2 o 26
1 In oduc ion
The ex en o gende bias in mo ali y in de eloping coun ies has been a esea ch
opic ha spa ked subs an ial con o e sies in he economic and demog aphic li e a-
u e since Ama ya Sen i s poin ed in 1989 o he millions o “missing women” in
Sou h and Eas Asia (Sen 1989). This li e a u e ocused la gely on calcula ing s ock
measu es o missing women. S ock measu es compa e he ac ual sex a io (males/
emales), using census in o ma ion in a coun y, wi h an expec ed sex a io ha
would occu in he absence o gende bias in mo ali y. An al e na i e me hod o he
s ock measu e poses he low measu e o missing women (Ande son and Ray 2010).
The low measu e compa es he a io o male- o- emale age-speci ic mo ali y a es
wi h an expec ed a io o male- o- emale age-speci ic mo ali y a es om a e e -
ence popula ion wi hou gende bias. Because low measu e es ima es can be be e
disagg ega ed by ime, age, and disease—whe eas he s ock measu e es ima es he
o al de ici in ali e women a one poin in ime— hey a e o g ea alue o a ge
policies and s imula e u u e esea ch. Howe e , hese wo measu es ha e e y di -
e en implica ions abou he o e all numbe o missing women and hei dis ibu-
ion by age and egion. This pape examines he low measu e’s sensi i i y o al e -
na i e assump ions and me hods and econciles he indings om he wo measu es.
Missing women es ima es based on he s ock measu e, and many mo e s udies on
gende inequali ies using mic o da a om indi idual coun ies o egions, ha e con-
e ged on wo key indings (e.g.,Sen 1989; Coale 1991; Coale and Banis e 1994;
Das Gup a 1987, 2005; Mu hi e al. 1995; Klasen and Wink 2002, 2003; Das Gup a
e al. 2009).1 Fi s , he p oblem o missing women is much mo e p e alen in China
and India han in Sub-Saha an A ica. Second, in China and India, missing women
a e la gely a ibu able o excess emale mo ali y be o e bi h and/o he i s ew
yea s o li e. The s ock es ima es amoun ed o a ound 100 million missing women
in 2000 in he mos a ec ed coun ies, i.e., China, India, Pakis an, Bangladesh, he
coun ies o Wes e n Asia, and some coun ies o No h A ica (Klasen and Wink
2003). Compa ed wi h a p e ious census ound in 1990, he numbe o missing
women had inc eased by abou 6 million be ween 1990 and 2000 bu allen as a
sha e o emales ali e ( om 6.5 o 5.7%) (Klasen and Wink 2003).2
1 The mic o-le el li e a u e has de eloped in o a massi e e idence base documen ing o Eas and Sou h
Asia no only he exis ence and de e minan s o gende disc imina ion in mo ali y, bu also in nume ous
o he heal h and beha io al ou comes, pa icula ly a bi h and he i s ea ly yea s o li e (e.g., As aw,
Klasen and Lamanna 2010; Jayachand an and Kuziemko 2011; Ba cellos e al. 2014; Jayachand an and
Rohini P, 2017; Ebe and Vollme 2022; Na h 2023). The impac policies o ac o s ha educe hese
gende imbalances ha e also been s udied (e.g., Anuk i i 2018; Tandel e al. 2023; Luo and Zou 2024;
Ja adeka and Saxena 2025).
2 A majo con o e sy abou hese s ock es ima es o missing emales a ose when Os e (2005) claimed
ha he “na u al” sex a io a bi h in some o hese coun ies was much highe due o he ac ha he
widesp ead p e alence o he hepa i is-B-ca ie s a us aised he sex a io a bi h. She claimed ha his
way, some 50–70% o missing emales could be accoun ed o ; i.e., hei absence was un ela ed o dis-
c imina ion. A e an in ense deba e and u he esea ch (Ab e aya 2009; Das Gup a 2005, 2006; Lin
and Luoh 2008), Os e e ac ed he inding in 2009 (Os e e al. 2010) and he p e ious es ima es o
a ound 100 million we e eins a ed. See Klasen (2008) o a e iew o his deba e.
Coun ing missing women: a econcilia ion o low ands ock… Page 3 o 26 82
Con a y o indings om he s ock measu e, es ima es o he annual low o missing
women o he yea 2000 in China, India, and Sub-Saha an A ica sugges ha excess
emale mo ali y is (a) as se e e among adul s as i is among child en in India and China,
(b) highe in Sub-Saha an A ica han in China and India, and (c) la gely due o excess
adul emale mo ali y in Sub-Saha an A ica (Ande son and Ray 2010; Wo ld Bank
2012). Fo China, India, and Sub-Saha an A ica oge he , excess mo ali y based on he
low measu e is also much la ge han implied by he exis ing li e a u e on s ock measu es.
The low measu e sugges s ha in a single yea , abou ou o i e million women died in
excess in India, China, and Sub-Saha an A ica alone. Fo he 1990 s, he s ock measu e
would be consis en wi h a low o abou one million missing women in a gi en yea .3
This pape discusses h ee key me hodological aspec s o he low measu e and
shows ha he low measu e esul s a e consis en wi h he s ock measu e esul s
when a se o di e en assump ions is made. The i s wo poin s ega d he e e ence
s anda d, which is used o simula e a gende -equal socie y, and he hi d poin ega ds
he a ailabili y o quali y mo ali y da a equi ed o he calcula ion o he low o
missing women. The ideal coun e ac ual o assessing he amoun o gende bias
in mo ali y would be a socie y in which males and emales a e ea ed equally and
in which he e a e no di e ences in sex-based beha io al pa e ns ac oss coun ies
wi h signi ican mo ali y implica ions (Coale 1991; Klasen and Wink 2002, 2003;
Ande son and Ray 2010). The e is no such socie y, pas o p esen , ha exhibi s hese
ea u es, so i is inhe en ly di icul o gene a e a e e ence s anda d ha is en i ely
beyond ep oach. Howe e , because i is so di icul o iden i y an adequa e e e ence
s anda d, i is impo an o unde s and he implica ions o he choices made. Th ough-
ou he pape , we e e o he low measu e based on he me hods and da a used in
Ande son and Ray (2010)— he seminal pape on he measu emen o he annual low
o missing women—and he e o e he numbe o missing women in 2000. The ci a-
ion coun o Ande son and Ray (2010) is 431, wi h he highes ci a ions pe yea
e y ecen ly in 2021 (38 ci a ions) and al eady 28 ci a ions his yea (2025).4 Ande -
son and Ray (2012, 2019) use he same me hods and es ima e excess emale mo al-
i y in Indian s a es and among unma ied women in de eloping coun ies.
Fi s , we e iew he e e ence popula ion o high-income coun ies in 2000 used in
low es ima es.5 Today’s high-income coun ies a e an en i onmen wi h ex emely
low o e all adul mo ali y and, he e o e, al eady small di e ences in male and emale
mo ali y a es esul in la ge changes in he male- o- emale mo ali y a e a io o
high-income coun ies. This is empi ically ele an as non-na u al causes o dea h o
3 No e ha lows o missing canno be calcula ed as di e ences in s ocks be ween ime pe iods; i.e., he
di e ence o 6 million in s ocks does no imply 600,000 missing women pe yea . This is due o he ac
ha some women in a pa icula age g oup ha ha e died in excess in a yea migh no coun as missing
in a s ock measu e a he nex census i in he ime in e al some men in ha age g oup ha e also died
due o high o e all mo ali y; i.e., o he ex en ha an excess emale dea h in a yea jus ad ances mo -
ali y be ween wo census pe iods, i would no add o he inc ease in he s ock o missing women a he
nex census.
4 Based on Google Schola and as o Decembe 2024.
5 High-income coun ies e e o he o me Wo ld Bank coun y classi ica ion “Es ablished Ma ke
Economies” as used by Ande son and Ray (2010), which includes Wes e n Eu ope, Canada, Uni ed
S a es, Aus alia, New Zealand, and Japan.
C.Ebe e al.
82 Page 4 o 26
consequences o beha io al pa e ns, such as acciden s, iolence, suicides, and smoking,
a ec males much mo e han emales. Second, we discuss ha disease en i onmen s do
no di e ac oss coun ies o males and emales in he same way, bu hey di e ac oss
sexes and coun ies. The e o e, con olling o di e ences in he disease en i onmen in
he e e ence s anda d equi es con olling o sex di e ences wi hin disease g oups ac oss
coun ies. Fo example, he male- o- emale mo ali y a io om AIDS is a ound 4 in high-
income coun ies in 2000 because AIDS mo ali y was con ined o male-domina ed high-
isk g oups (gay men and IV d ug use s), whe eas in Sub-Saha an A ica AIDS sp ead
om he beginning o he en i e popula ion, wi h he e osexual ansmission playing a key
ole, and AIDS mo ali y is mo e simila o males and emales. Thus, using oday’s high-
income coun ies as a e e ence popula ion o AIDS mo ali y sugges s massi e excess
emale mo ali y in adul hood in coun ies wi h high a es o AIDS mo ali y. The Wo ld
De elopmen Repo 2012, which uses he low measu e app oach, indeed shows pa icu-
la ly high a es and inc eases in excess emale mo ali y in AIDS-a ec ed coun ies o
Sub-Saha an A ica (Wo ld Bank 2012). The wo poin s on he low mo ali y en i on-
men and he disease co ec ion a e ela ed, bu each has a sepa a e e ec on he es ima es
o missing women and is he e o e also discussed sepa a ely. In a hi d poin , we highligh
a gene al conce n ha unde lies he use o mo ali y da a in Sub-Saha an A ica, which is
ha eliable i al egis a ion da a is only a ailable in one o 46 coun ies. Fo he emain-
ing coun ies, mo ali y da a mus be es ima ed and, he e o e, he low o missing women
is inhe en ly imp ecise.
We use a a ie y o al e na i e e e ence s anda ds o sex-speci ic mo ali y and
ind pa e ns in he low o missing women o he yea 2000 ha a e consis en wi h
p e ious indings on he s ock o missing women. We ind ha excess emale mo al-
i y is much mo e se ious in India and China han in Sub-Saha an A ica. In China and
India, excess emale mo ali y is la gely d i en by gende bias p e-bi h and du ing he
i s ew yea s o li e. The magni ude o he p oblem is also much educed, wi h ou
es ima es sugges ing ha 1.5–1.8 million women die in excess in he h ee egions in
2000. This numbe is subs an ial and wo ying bu below he numbe o i e million
emales pe yea es ima ed in Ande son and Ray (2010) o 3.3 million unde he age
o 60 by he Wo ld Bank (2012). Ou es ima e is also lowe , bu o a lesse ex en , han
he low measu e es ima e by Bongaa s and Guilmo o (2015), who show ends in
excess emale mo ali y om 1970 o 2050 using da a om 93 coun ies wi h di e -
en mo ali y le els o c ea e a e e ence popula ion.6 Bongaa s and Guilmo o (2015)
6 The es ima ions o Bongaa s and Guilmo o (2015) amoun o excess emale dea hs a e bi h o 0.54
million in China and 0.82 million in India in 2000. They p edic he e e ence sex a io o mo ali y
based on a linea eg ession o he log sex a io on li e expec ancy o combina ions o di e en age
g oups and yea s (304 sepa a e eg essions). The coun ies included in he compu a ion o he e e ence
popula ion a e Alge ia, Angola, A gen ina, Aus alia, Aus ia, Belgium, Benin, Boli ia, B azil, Bu kina
Faso, Bu undi, Cambodia, Came oon, Canada, Chad, Chile, Colombia, Cô e d’I oi e, Cuba, Democ a ic
Republic o he Congo, Denma k, Dominican Republic, Ecuado , Egyp , El Sal ado , E i ea, E hiopia,
Finland, F ance, Ge many, Ghana, G eece, Gua emala, Guinea, Hai i, Hondu as, Indonesia, I an, I aq,
Is ael, I aly, Japan, Jo dan, Kazakhs an, Kenya, Ky gyzs an, Laos, Libya, Madagasca , Malawi, Mali,
Mexico, Mo occo, Mozambique, Myanma , Ne he lands, Nica agua, Nige , Nige ia, No h Ko ea, Papua
New Guinea, Pa aguay, Pe u, Philippines, Po ugal, Rwanda, Senegal, Se bia, Sie a Leone, Somalia,
Sou h A ica, Sou h Sudan, Spain, S i Lanka, Sudan, Sweden, Swi ze land, Sy ia, Tajikis an, Thailand,
Togo, Tunisia, Tu key, Tu kmenis an, Uganda, he UK, Tanzania, Uni ed S a es, Uzbekis an, Venezuela,
Yemen, Zambia, and Zimbabwe.
Coun ing missing women: a econcilia ion o low ands ock… Page 5 o 26 82
do no p o ide low es ima es by age, so compa isons o age pa e ns canno be made.
Fu he , ou calcula ions con i m se e e gende bias in mo ali y among adul s in Sub-
Saha an A ica, likely linked o ele a ed mo ali y o women due o AIDS, al hough i
is o smalle magni ude han he o iginal low es ima es.
These esul s a e impo an because missing women es ima es guide policy-mak-
ing. As e iden om he Wo ld De elopmen Repo on “Gende Equali y and De el-
opmen ,” he low o missing women poses a ele an me ic o highligh g ie ances in
gende equali y in de eloping coun ies as hey allow o compa isons by age, disease,
and o e ime. The Wo ld De elopmen Repo concludes, “Excess emale mo ali y
is slowly shi ing om ea ly childhood in Sou h Asia o adul hood in Sub-Saha an
A ica, declining in all low-income coun ies excep in Sub-Saha an A ica” (Wo ld
Bank 2012, p. 77). The epo u he poin s ou ha some o he esul s we e no
de ec ed byo a e e en in con as wi h ea lie s ock measu e es ima es: “Less well
known is ha excess emale mo ali y is a con inuing phenomenon beyond childhood
and a g owing p oblem in Sub-Saha an A ica. […] While missing gi ls a bi h a e
indeed concen a ed in India and China, consis en wi h he ea lie discussion, excess
emale mo ali y a e bi h is highes in Sub-Saha an A ica, he only egion whe e
he numbe s a e going up o e ime” (Wo ld Bank 2012, pp. 118–120). While ou low
es ima es di e om he p e ious ones in he size and age pa e ns o missing women,
hey do ein o ce ha he numbe o missing women aged 15 o 25 in Sub-Saha an
A ica is la ge and should be a subjec o u u e esea ch.
This pape is o ganized as ollows. In Sec .2, we p esen he concep ual ame-
wo k o he low measu e o missing women. In Sec .3, we discuss he p oblem o
low-mo ali y e e ence popula ions and he sex a io as a unc ional o m o he e -
e ence s anda d, and p esen es ima es o missing women using al e na i e e e ence
popula ions and unc ional o ms. In Sec .4, we discuss he ele ance o sex-speci ic
di e ences wi hin diseases ac oss coun ies and p esen es ima es o missing women
using an al e na i e e e ence s anda d o AIDS. In Sec .5, we d aw a en ion o he
issue o lacking mo ali y da a in Sub-Saha an A ica. Sec ion6 concludes.
2 The low measu e o missing women
Using he low measu e, excess emale mo ali y (
EFM
) o age g oup
a
can be cal-
cula ed he ollowing way:
MR (a)
and
MRm(a)
e e o he ac ual mo ali y a es o emales and males
in age g oup
(a)
, espec i ely.
MR
(a
)
and
MRm(a)
desc ibe he mo ali y a es
o emales and males in age g oup
(a)
in he e e ence popula ion, espec i ely.
MRm
(a)∕
MR
(a
)
is he male- o- emale sex a io o he e e ence popula ion and is
EFM
(a)=
(
MR (a)−
(
MRm(a)
MR
m
(a)∕
MR
(a)
))
∗Pop (a
)
C.Ebe e al.
82 Page 6 o 26
used as a co ec ion ac o o scale he ac ual male mo ali y a e
MRm(a)
o accoun
o , o example, di e ences in mo ali y om isky beha io o speci ic diseases
ac oss sexes. The scaled
MRm(a)
p o ides he e e ence mo ali y a e o emales in
age g oup
a
which is sub ac ed om he ac ual mo ali y a e o emales o ecei e
he popula ion sha e o missing women in age g oup
a
. Mul iplying he popula ion
sha e o missing women in age g oup
a
wi h he numbe o women in age g oup
a
,
Pop (a),
esul s in he o al low o missing women in age g oup
a
,
EFM(a)
. Based
on Eq. (1), he essen ial ing edien s o calcula e excess emale mo ali y a e he
ac ual age-speci ic mo ali y a es and he co ec ion ac o consis ing o he age-
speci ic male- o- emale mo ali y a e a io o he e e ence popula ion. As a e e -
ence popula ion, he low measu e o missing women uses high-income coun ies in
2000. The e e ence popula ion and he unc ional o m o he co ec ion ac o , i.e.,
di iding by he sex a io in mo ali y, oge he build he e e ence s anda d.
Table1 p esen s he mo ali y a es o adul s aged 15 o 59 by sex o Sub-Saha-
an A ica, India, and China in 2000 as ac ual mo ali y a es (
MR (15 −59)
and
MRm(15 −59)
) and o high-income coun ies in 2000 as e e ence mo ali y a es
(
MRm
(15 −59
)
and
MR
(15 −59
)
). The ac ual sex a ios o mo ali y a es ange
be ween 1.13 and 1.46. Adul males hus die a 13–46% highe a es han adul
emales. The sex a io in high-income coun ies in 2000 is abou 1.92. Thus, adul
males die a 92% highe a es han adul emales. This sex a io in high-income coun-
ies cons i u es he key componen o he low measu e’s e e ence s anda d and he
es ima es o missing women will be highly sensi i e o his e e ence s anda d.
Following Eq.(1) and using high-income coun ies in 2000 as a e e ence popula-
ion, Table2p esen s he low o missing women in Sub-Saha an A ica, India, and
China in 2000 (Ande son and Ray 2010).7 Wi h his e e ence s anda d, he p oblem
o missing women as a sha e o he emale popula ion is wo s in Sub-Saha an A ica.
0.47% o emales died in excess in 2000 in Sub-Saha an A ica, compa ed o 0.35%
in India and 0.28% in China. Table2 u he illus a es ha in India and China, he e
is a massi e p oblem o missing women among adul s ha is la ge han among chil-
d en. A las , adding he o al numbe o missing women ac oss Sub-Saha an A ica,
India, and China esul s in 5 million emales ha wen missing in 1 yea .
The ex en o which hese excess emale dea hs a e en i ely a ibu able o di ec
gende disc imina ion, a he han o he causes, has go en ca e ul a en ion: “i is
possible ha ce ain aspec s o po e y can c ea e ‘unin ended’ gende biases ela-
i e o he de eloped-coun y benchma k. The accoun ing me hodology ha we ol-
low is en i ely silen on ma e s o in e p e a ion. The case o (o agains ) lack o
simila ca e has o be made sepa a ely” (Ande son and Ray 2010, p. 1292). Also,
his pape is silen abou he causes unde lying a ce ain numbe o missing women;
ins ead, i discusses he es ima ion o he numbe i sel .
7 Fo example, in Sub-Saha an A ica, he mo ali y a e o emales aged 1–5 is 1.39%, while he male
one is 1.27% (Ande son and Ray 2010, Table1). Since in high-income coun ies he male– emale sex
a io is 1.25 in ha age g oup (Ande son and Ray 2010, Table1), he numbe o missing emales aged
1–5 in Sub-Saha an A ica is (0.0139 − (0.0127/1.25))×42m. = 0.16 million as shown in Table1.
Coun ing missing women: a econcilia ion o low ands ock… Page 7 o 26 82
3 Low mo ali y e e ence popula ions and hesex a io
asa unc ional o m o he e e ence s anda d
3.1 The p oblem
The low measu e o missing women employs male- o- emale age-speci ic mo al-
i y a e a ios ha p e ail in high-income coun ies in 2000 as a e e ence s anda d.
The main sho coming o his e e ence s anda d is he e y low o e all mo ali y
in high-income coun ies. In e y low mo ali y en i onmen s, e en he smalles
di e ences in he numbe s o dea hs ac oss sexes esul in e y high mo ali y sex
a ios and he e o e low emale e e ence mo ali y a es and high lows o missing
women. In he adul age g oup o 15- o 29-yea -olds, which accoun s o mos o he
es ima ed low o missing women in Sub-Saha an A ica (see Table2), he emale
mo ali y a e in high-income coun ies in 2000 is 0.42 pe housand, whe eas ha
o men is sligh ly highe a 1.08 pe housand, esul ing in a sex a io o 2.6; i.e.,
he p obabili y o men dying in his young age g oup is 160% highe han he co e-
sponding likelihood o women. By con as , he emale mo ali y a e a io o 25- o
29-yea -olds in Sub-Saha an A ica is 14.8, and ha o males is 10.8.
In Fig.1, we simula e how he mo ali y sex a io o 25- o 29-yea -olds in
he e e ence popula ion would all as he mo ali y le el inc eases om e y
low mo ali y le els p e ailing in high-income coun ies o high le els p e ail-
ing in Sub-Saha an A ica, while he di e ence in mo ali y a es be ween sexes
emains cons an . Figu e1 also shows he numbe o missing women associa ed
wi h hese mo ali y le els and sex a ios. Inc easing he emale mo ali y a e o
25- o 29-yea -olds only sligh ly om 0.42 o 2.42 educes he sex a io by hal
om 2.6 o 1.3, and he numbe o missing women alls om 258,118 o 152,434,
a di e ence o o e 100,000 jus in he age g oup o 25- o 29-yea -olds alone.
Fu he inc eases in he mo ali y le el ha e much smalle impac s on he numbe
o missing women because he mo ali y sex a io is inc easingly less sensi i e as
mo ali y le els ise. Inc easing he emale mo ali y a e o 25- o 29-yea -olds
all he way o he le el p e ailing in Sub-Saha an A ica, which is 14.8, educes
he numbe o missing women o below 110,000.8 This sensi i i y o he numbe
o missing women is pa icula o he low measu e because he s ock measu e
does no ely on he numbe o dea hs bu on he numbe o males and emales
su i ing, which make up, o cou se, much la ge numbe s.
The abo e calcula ions illus a e he sensi i i y o he numbe o missing
women wi h espec o he mo ali y le el o 25- o 29-yea -olds only, which
is he age g oup wi h he highes mo ali y sex a io. Figu e2 depic s he male-
o- emale mo ali y a ios o all age g oups in high-income coun ies in 2000.
The a ios a e pa icula ly high – abo e 2 – in he age g oups be ween 15 and
35. Be ween he ages o 35 and 55, he mo ali y a ios a e somewha smalle ,
8 A high mo ali y le els, also he di e ence be ween he male and emale mo ali y a e is o en la ge .
In Sub-Saha an A ica, he di e ence is 3.9. I we we e o use ha di e ence ins ead, o inc easing he
di e ence om 0.66 o 3.9 as we inc ease he mo ali y le el om 0.42 o 14.8, he implica ions o his
simula ion exe cise would no change.
C.Ebe e al.
82 Page 8 o 26
a e aging a ound 1.86. A inal peak a ises be ween he ages o 55 and 70, wi h
mo ali y a e a ios o e 1.94. Fo olde ages, he a io alls sha ply, bu always
emains abo e 1.
The i s peak be ween 15 and 35 is la gely a esul o he ole o inju ies
which pa icula ly d i es up male mo ali y a es in hese young age g oups; in
he demog aphic li e a u e, his has some imes been e e ed o as he “ es os e -
one spike” linked o inju ies ela ed o he combina ion o alcohol abuse, dange -
ous a ic beha io , and iolence among young men (Kalben 2000). Because o
he o e all low mo ali y among 15- o 29-yea -olds in high-income coun ies,
dea hs om inju ies a e pa icula ly p ominen . In his young age g oup, inju ies
make up 71% o all dea hs o males (and 51% o emales), and he sex a io o
mo ali y om inju ies is 3.7, whe eas i is 1.6 o all o he dea hs. The e o e, he
e e ence s anda d is hea ily a ec ed by his disp opo iona e impo ance o inju-
ies in adul mo ali y o males in high-income coun ies.
The second peak is o a la ge ex en caused by he mo ali y consequences o
high a es o smoking among adul s. As shown in de ail by C immins e al. (2011),
wha ma e s o mo ali y a e smoking a es wo o ou decades p io , i.e., smoking
Table 1 Adul mo ali y a es
(15–59) by sex in 2000
No es: Adul mo ali y a es e e o he likelihood o ha ing died by
age 60 i one was ali e a age 15. Male e e s o he adul (15–59)
mo ali y a e among males. Female e e s o he adul (15–59) mo -
ali y a e among emales. Sex a io e e s o he a io o columns
(1) and (2), i.e., he male- o- emale mo ali y a e a ios. To de i e
mo ali y a es in Sub-Saha an A ica and high-income coun ies,
popula ion (15–59 yea s old) weigh ed a e ages o coun y speci ic
adul mo ali y a es we e calcula ed (see supplemen a y ma e ials).
Sub-Saha an A ican coun ies include: Alge ia, Angola, Benin,
Bo swana, Bu kina Faso, Bu undi, Came oon, Cape Ve de, CAR,
Chad, Como os, Congo, Co e d’I oi e, DRC, Equa o ial Guinea,
E i ea, E hiopia, Gabon, Gambia, Ghana, Guinea, Guinea-Bissau,
Kenya, Leso ho, Libe ia, Madagasca , Malawi, Mali, Mau i ania,
Mau i ius, Mozambique, Namibia, Nige , Nige ia, Rwanda, Sao
Tome and P incipe, Senegal, Seychelles, Sie a Leone, Sou h A ica,
Swaziland, Tanzania, Togo, Uganda, Zambia, Zimbabwe. High-
income coun ies include: Aus alia, Aus ia, Belgium, Canada,
Denma k, Finland, F ance, Ge many, Iceland, I eland, I aly, Japan,
Luxembou g, Ne he lands, New Zealand, No way, Po ugal, Spain,
Sweden,Swi ze land, Uni ed Kingdom, Uni ed S a es
Sou ces: Coun y speci ic adul mo ali y a es in 2000 we e aken
om Wo ld Heal h O ganiza ion (2001). Popula ion numbe s we e
aken om Uni ed Na ions, Depa men o Economic and Social
A ai s, Popula ion Di ision (2000)
Male Female Sex a io
(1) (2) (3)
Sub-Saha an A ica 0.511 0.454 1.125
India 0.287 0.213 1.347
China 0.161 0.110 1.464
High-income coun ies 0.125 0.065 1.918
Coun ing missing women: a econcilia ion o low ands ock… Page 15 o 26 82
mo ali y om AIDS han men. The s a k di e ence in sex a ios o AIDS mo al-
i y ac oss egions eme ges because AIDS incidence and a ali ies in high-income
coun ies a e mos ly d i en by gay men and IV d ug use s (also p edominan ly
men), whe eas he e osexual ansmission plays a signi ican ole in Sub-Saha an
A ica (Ainswo h and O e 1997). Sex-speci ic di e ences in mo ali y ac oss
coun ies also exis , as p e iously discussed, o dea hs om isky beha io s.
Howe e , whe eas sex-speci ic di e ences in isky beha io a e likely simila in
Sub-Saha an A ican coun ies and high-income coun ies, he pa e ns o o he
diseases, such as AIDS, e iden ly di e ge o a la ge ex en .
To accoun o di e ences in disease en i onmen s, he o iginal low meas-
u e neglec s sex-speci ic di e ences and builds up he numbe o missing women
disease g oup by disease g oup. Speci ically, he sex a io o age-speci ic mo al-
i y a es o each disease g oup in high-income coun ies is used as he e e ence
Table 3 Numbe o missing women (in 1000 s) by age g oup using al e na i e e e ence s anda d
No es: Ra io e e s o he unc ional o m o he e e ence s anda d ha di ides he male mo ali y a e
in he egion o in e es by he sex a io o mo ali y a es in he e e ence popula ion. Di . e e s o he
unc ional o m o he e e ence s anda d ha sub ac s he di e ence in sex-speci ic e e ence popula ion
mo ali y a es om he male mo ali y a e in he egion o in e es . RC e e s o high-income coun ies
in 2000; MLT a e P ince on Model Li e Tables Wes . Numbe s do no sum o o al because o ounding
Sou ces: Popula ion and dea h numbe s by age o Sub-Saha an A ica, India, and China, and he e e -
ence popula ion o high-income coun ies in 2000 we e aken om he Supplemen a y Ma e ials 2 o
Ande son and Ray (2010). Mo ali y a es o P ince on Model Li e Tables Wes we e aken om Coale
e al. (1983). To de e mine he co ec mo ali y le el in Coale e al. (1983), li e expec ancy a bi h o
Sub-Saha an A ica, India, and China in 2000 was aken om Uni ed Na ions, Depa men o Economic
and Social A ai s, Popula ion Di ision (2019)
Sub-Saha an A ica India China
RC MLT RC MLT RC MLT RC MLT RC MLT RC MLT
Ra io Ra io Di Di Ra io Ra io Di Di Ra io Ra io Di Di
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
A bi h 0 0 0 0 184 184 184 184 644 644 644 644
0–1 32 22 − 235 13 146 181 8 156 109 148 65 119
1–4 160 55 55 56 164 140 121 141 23 29 13 20
5–14 70 5 18 2 93 71 53 69 2 12 − 9 4
15–29 578 241 279 237 258 111 127 113 24 − 9 − 9 − 37
30–44 345 − 3 − 56 − 51 93 − 70 − 63 − 80 73 − 2 53 − 16
45–59 84 − 55 − 144 − 93 120 − 39 − 114 − 75 89 − 9 15 − 26
60–79 213 − 40 − 18 − 40 541 76 103 41 490 27 170 12
80 + 44 − 10 10 − 8 113 21 84 39 272 64 165 64
To al 1526 213 − 91 116 1712 674 503 590 1727 903 1107 783
% o
emales
0.47 0.07 − 0.03 0.04 0.35 0.14 0.10 0.12 0.28 0.14 0.18 0.13
C.Ebe e al.
82 Page 16 o 26
s anda d, and he age- and disease-speci ic lows a e hen summed up o e all ages
and diseases o achie e he o al numbe o missing women in a yea .19
Building up he low o missing women disease by disease only has mino impli-
ca ions o he es ima ed o al numbe o missing women. In Sub-Saha an A ica,
he annual low alls om 1.526 million o 1.385 million, in India om 1.712 o
1.637 million, and in China om 1.727 o 1.592 million (Ande son and Ray 2010).
The key indings o he o e all high magni ude (4.6 million pe yea ), he p eponde -
ance among adul s, and he la ge sha e o missing women in A ica a e no a ec ed
signi ican ly by his disease co ec ion.
The e e ence s anda d used in his disease co ec ion p esumes a coun e ac ual
ha wi hou gende bias, AIDS in Sub-Saha an A ica would ha e also been p ima -
ily a disease a ec ing gay men and IV d ug use s and sugges s ha 58% o all adul
missing women in Sub-Saha an A ica die om AIDS (see Table4, column (3)).
This coun e ac ual is likely alse. AIDS in A ica has been a disease ha has sp ead
mos ly h ough he e osexual in e cou se (and mo he o child ansmission) due o
he ea ly eme gence o he disease in A ica and i s sp ead among he e osexual cou-
ples long be o e i was iden i ied as a disease, oge he wi h a e y high incidence o
STDs (Ili e 2006; Ainswo h and O e 1997; Os e 2005). I is likely o be ue ha
he pa icula ly high incidence and a ali y o AIDS among young women is o some
ex en due o gende disc imina ion (e.g., unwan ed sexual in e cou se wi h olde
men), bu his is no he main eason why AIDS is a disease a ec ing mos ly he -
e osexual people in Sub-Saha an A ica ac oss adul age g oups. Assuming he a io
om high-income coun ies as a e e ence s anda d likely o e s a es he ele ance o
AIDS in causing missing women conside ably.
To es he sensi i i y in he numbe o missing women o al e na i e assump-
ions o an AIDS-speci ic e e ence s anda d in en i onmen s wi h a p edominan ly
he e osexual ansmission pa h, we assume a sex a io o mo ali y in high-income
coun ies ha weighs mo ali y a es by he sha e o he e o- and homosexual popu-
la ions. Fo example, he sex a io o HIV among he e osexual men (84.1%) and
women (86.9%) as well as homosexual men and women (de ined as hose who a e
no he e osexual) in he US, as one o he es ablished ma ke economies, is a ound 1
(Lansky e al. 2015; Cen e s o Disease Con ol and P e en ion 2019).20 This num-
be is much smalle han he ac ual male- o- emale mo ali y a io om AIDS o 4
p e ailing in mos adul age g oups o high-income coun ies in 2000. The sex a io
o AIDS mo ali y o 1 would s ill allow o high emale mo ali y om AIDS in
he 15–29 age g oup—in which gende disc imina ion in he ansmission o AIDS
19 When emale dea hs om a disease we e less han 100 in he conside ed age g oup, Ande son and
Ray (2010) use a e age disease g oup-speci ic e e ence mo ali y a ios, a he han he mo ali y a io
o ha one disease. When dea hs we e essen ially ze o in he e e ence popula ion, such as STDs among
child en in de eloped coun ies, a e e ence sex a io o 1:1 was used (Ande son and Ray 2010, p.1279).
20 Speci ically, he sex a io is calcula ed as:
sex a io
=
(sha ehe e o.men∗HIVcasespe 100,000he e o.men)+(sha ehomo.men∗HIVcasespe 100,000homo.men)
(sha ehe e o.women∗HIVcasespe 100,000he e o.women)+(sha ehomo.women∗HIVcasespe 100,000he e o.women)
.
Coun ing missing women: a econcilia ion o low ands ock… Page 17 o 26 82
h ough o ced sexual in e cou se likely ma e s—as he male- o- emale a io o
dea hs om AIDS among 15- o 29-yea -olds is only 0.53 in Sub-Saha an A ica.21
Table4p esen s he low o missing women om AIDS using a sex a io o mo -
ali y om AIDS o 1 in columns (4) and (5) and a sex a io o 1.2 as a mo e con-
se a i e e e ence s anda d o AIDS in columns (6) and (7). The low o missing
women om AIDS educes by abou 400,000 excess emale dea hs, om 611,000 o
19,000–171,000, essen ially all acc uing in he 15–29 age g oup. Fu he , he sha e
o adul women ha go missing due o AIDS educes om 58% o a mo e ealis ic
pe cen age sha e o 4 o 28%.
4.2 Ma e nal mo ali y, inju ies, andsuicide
The e a e nume ous o he diseases in addi ion o AIDS ha equi e a en ion when
he e e ence popula ion is no d awn om a compa able disease en i onmen . Some
diseases a e pa icula ly a e in high-income coun ies, which ende s he sex a io
o mo ali y p e ailing in high-income coun ies an o e all unsui able e e ence
s anda d o ha pa icula disease.
As he e is no such hing as pa e nal mo ali y ha can be used o de i e expec ed
ma e nal mo ali y, he low measu e o missing women assumes ha any ma e nal
dea h abo e he a e p e ailing in high-income coun ies is an excess dea h. Because
ma e nal mo ali y is a e in high-income coun ies, ma e nal mo ali y among 15-
o 45-yea -olds accoun s o he second mos excess emale dea hs in Sub-Saha an
A ica (a e AIDS, which would change i he al e na i e e e ence s anda ds o
AIDS we e used) and o mos excess emale dea hs in India. Hence, he assump ion
o all ma e nal dea hs being excess emale dea hs is a de e mining ac o o he o al
numbe o missing women and he dis ibu ion o missing women ac oss age g oups
in Sub-Saha an A ica and India.
Ma e nal mo ali y is ypically ela ed o he poo o e all heal h condi ions and
se ices p e ailing in many de eloping coun ies. Figu e3 p esen s he ela ionship
be ween child mo ali y a es o bo h sexes, which is an indica o o o e all heal h
condi ions in a coun y, and he ma e nal mo ali y a io. Child and ma e nal mo al-
i y show a close i , indica ing ha ma e nal mo ali y goes hand-in-hand wi h poo
o e all heal h. I Sub-Saha an A ican coun ies we e ou lie s in his ela ionship,
hen his migh be due o gende inequali y in heal h access ha goes beyond poo
o e all heal h condi ions. The dashed lines in Fig.3 p esen he p edic ed ma e -
nal mo ali y a ios based on a simple linea eg ession model o ma e nal mo ali y
on child mo ali y and a dummy o Sub-Saha an A ica. The esul s sugges ha
on a e age, abou 11–16% o he ma e nal mo ali y in Sub-Saha an A ica is no
21 While pa ly his high mo ali y could indeed be due o disc imina ion linked o unwan ed sexual
in e cou se o young women wi h olde men, pa o he high emale mo ali y could also be due o
ea lie olun a y sexual ac i i ies o emales (and hus ea lie AIDS isks) which is obse able in mos
coun ies o he wo ld, including high-income coun ies; see Wo ld Bank (2012) o a discussion.
C.Ebe e al.
82 Page 18 o 26
explained by poo heal h condi ions and migh be due o gende bias.22 Thus, a small
numbe o dea hs om ma e nal mo ali y—a ound 22,000 o 32,000 a he han all
200,000 ma e nal dea hs—migh indeed be due o gende inequali y.
Ano he no ewo hy cause o dea h is inju ies, speci ically suicides and i es,
o which di e ences in he dis ibu ion o dea hs by sex po en ially accoun o
a subs an ial numbe o missing women. The sex a io o dea hs om suicides in
India is ela i ely balanced and likely d i en by he economic ha dship o amilies
(especially in u al a eas), whe eas dea hs om suicides in high-income coun ies
Table 4 Numbe o missing
women om AIDS (in 1000 s)
by age g oup using al e na i e
e e ence s anda ds
No es: Column (1) p esen s he o al numbe o missing women
when calcula ed disease g oup by disease g oup as p esen ed in
Table 6 o Ande son and Ray (2010). Columns (2) o (7) e e o
di e en e e ence s anda ds ha we e applied in he es ima ion o
missing women om AIDS. To calcula e he % o o al in columns
(5) and (7), he numbe s in columns (4) and (5) we e di ided by
he new o al numbe o missing women ha we e co ec ed o he
numbe o missing women om AIDS, i.e., 180/(406-(277–180))
= 0.58. Figu es in columns (1), (2), (4), and (6) a e ounded o he
nea es housand. Pe cen age sha es in columns (3), (5), and (7) a e
based on numbe s ounded o nea es housand in columns (1), (2),
(4), and (6). Numbe s do no sum o o al because o ounding
Sou ces: Ande son and Ray (2010) Table6 o column (1). Popula-
ion numbe s and dea h numbe s om AIDS by age o Sub-Saha an
A ica and he e e ence popula ion o high-income coun ies in
2000 we e aken om he Supplemen a y Ma e ials 2 o Ande son
and Ray (2010)
Sub-Saha an A ica: Flow o missing women
To al F om AIDS
Re e ence:
sex a io,
high-income
coun ies
Me hod:
disease-by-
disease g oup
Re e ence:
sex a io,
high-income
coun ies
Re e ence: sex
a io = 1
Re e ence: sex
a io = 1.2
(1) (2) (3) (4) (5) (6) (7)
Age No No % o o al No % o o al No % o o al
15–29 406 277 68% 180 58% 213 62%
30–44 278 240 86% − 91 – − 8 –
45–59 134 78 58% − 58 – − 29 –
60–79 217 15 7% − 11 – − 5 –
80 + 25 0 – 0 – 0 –
To al 1,060 611 58% 19 4% 171 28%
22 11–16% is he con idence in e al o he mean o he Sub-Saha an A ica ixed e ec (81 dea hs)
di ided by he p edic ed ma e nal mo ali y a io (mean = 13.5%, CI = 10.8–16.2%).
Coun ing missing women: a econcilia ion o low ands ock… Page 19 o 26 82
a e male-biased and d i en by psychia ic condi ions (e.g.,Adi janyee 1986; Hi oeh
e al. 2001; Mo ensen e al. 2000; No den o e al. 1993; Qin e al. 2000). Simila ly,
dea hs om i es a e hea ily male-biased in high-income coun ies (sex a io o
adul mo ali y (15–59) o 2.1), whe eas in India, i es a lic ela i ely mo e women
(sex a io o adul mo ali y (15–59) o 0.44). An es ima ed 106,000 ou o 163,000
i e- ela ed dea hs in India in 2001 acc ued o women (Sangha i e al. 2009; Bhalla
and Sangha i 2020). I was u he es ima ed ha mo e women died om i es han
om gi ing bi h and a a a e 20 imes mo e han in he es o he wo ld. These
i e- ela ed dea hs a e said o be due o he use o open i es, candles, and ke osene
du ing domes ic cho es as well as sel -immola ion and domes ic iolence (Sangha i
e al. 2009). Thus, a conside able sha e o i e- ela ed missing women is likely due
o gende bias.23 Ye , he low o adul (> 15 yea s) missing women om i es using
he sex a io om high-income coun ies as a e e ence s anda d is 72,000 in India,
which makes up o e 50% o missing women om unin en ional inju ies and 80%
o all i e- ela ed dea hs o adul women would ha e o be due o gende disc imina-
ion, which seems unlikely.
The discussion illus a es ha when using he sex a io o mo ali y p e ailing in
oday’s high-income coun ies as a e e ence s anda d in he es ima ion o missing
women, disease co ec ions mus add ess he di e ences in he sex-speci ic disease
en i onmen s be ween high-income coun ies and China, India, and Sub-Saha an
A ica. Di e ences in disease en i onmen s may be based on di e en s uc u es
wi hin disease g oups (as in AIDS o inju ies) o can be la gely a esul o he o e -
all wo se disease en i onmen in de eloping coun ies (as is he case wi h ma e nal
mo ali y).
O e all, he necessi y o a de ailed disease co ec ion depends on he chosen
e e ence popula ion. The mo e app op ia e he gene al e e ence popula ion is, he
less such a disease co ec ion is equi ed. Fo example, es ima es using he P ince-
on Model Li e Tables a e much mo e simila o he wi hin-disease co ec ed low
o missing women using high-income coun ies as a e e ence s anda d al eady
because he P ince on Model Li e Tables o e e e ence mo ali y a es o mo ali y
en i onmen s ha a e simila o he coun y o egion o in e es .
5 A ailabili y andquali y o mo ali y da a omSub‑Saha an A ica
Pu ing he me hodological ques ion o app op ia e e e ence s anda ds aside, an
impo an eason why low es ima es o missing women ha e no been p oduced
be o e is he lack o eliable da a on age-speci ic mo ali y a es by sex in many
de eloping coun ies. Disease-speci ic da a a e e en ha de o come by as hey
equi e a comple e i al egis a ion sys em.
23 Due o dea h om in e pe sonal iolence and women would also be less likely o conduc household
cho es i he di ision o domes ic and wage wo k was equal ac oss sexes.
C.Ebe e al.
82 Page 20 o 26
Ou analysis ollows p e ious es ima ions o he low o missing women and
uses mo ali y da a o he yea 2000 om he Wo ld Heal h O ganiza ion’s 2002
Global Bu den o Disease es ima es (Ande son and Ray 2010; Wo ld Bank 2012).24
A de ailed desc ip ion o da a sou ces and me hods used in he de i a ion o he
WHO mo ali y da a is p o ided in Ma he s e al. (2004; see ex , Table4, Annex
Table 6, and Annex Table7). While he majo i y o coun ies ha e ai ly ecen
da a on in an and child mo ali y om DHS and ela ed su eys ( he modal ime
pe iod being 1995–1999; see Ma he s e al. Table3), da a o es ima e adul mo al-
i y a e o en lacking. In Sub-Saha an A ica, adul mo ali y es ima es o 2000 a e
based on obse a ions om only 4 ou o 46 coun ies, wo each o he ea ly and
he la e 1990 s (see Ma he s e al. 2004, Table2). In ac , eliable i al egis a ion
da a a e a ailable o only one coun y in Sub-Saha an A ica, which is he island
o Mau i ius (see Ma he s e al. 2004, Annex Table6). Al hough China and India
also lack ull i al egis a ion sys ems, in China, age-speci ic dea hs could be es i-
ma ed om a sample i al egis a ion and he 2000 census, and in India, da a om
Fig. 3 C oss-coun y ela ionship be ween child mo ali y (x-axis) and ma e nal mo ali y (y-axis).
No es: Child mo ali y is measu ed as he p obabili y o dying be ween bi h and exac age 5, exp essed
pe 1000 li e bi hs, while he ma e nal mo ali y a io (MMR) is he numbe o ma e nal dea hs pe
100,000 li e bi hs. P edic ed ma e nal mo ali y a io is based on a linea p edic ion o ma e nal mo al-
i y on child mo ali y and a Sub-Saha an A ica dummy (bo h p edic ions esul om he same model).
Sou ces: Uni ed Na ions In e -agency G oup o Child Mo ali y Es ima ion (2021) o child mo ali y in
2000. Hogan e al. (2010) o ma e nal mo ali y a io in 2000. Uni ed Na ions, Depa men o Economic
and Social A ai s, Popula ion Di ision (2019) o assignmen o coun ies o SDG egion Sub-Saha an
A ica and o es o wo ld
24 We di ec ly use he age- and disease-speci ic mo ali y da a o Ande son and Ray (2010) as p o ided
in hei supplemen a y ma e ials.
Coun ing missing women: a econcilia ion o low ands ock… Page 21 o 26 82
a p ojec i al egis a ion sys em we e a ailable (see Ma he s e al. 2004, pp. 10/11
and Annex Table6).
Since he Global Bu den o Disease es ima es in 2002, he e ha e been mul iple
e isions and upda es o he WHO li e ables, he la es upda e being om Decem-
be 2019. Howe e , he a ailabili y o quali y i al egis a ion da a has no changed
since 2002.25 Al hough a ew mo e coun ies, o example, Cabo Ve de and Sou h
A ica, now collec and epo i al egis a ion da a o he Wo ld Heal h O gani-
za ion, hey we e excluded due o insu icien quali y (Wo ld Heal h O ganiza ion,
2020 a,b). In consequence, e en he mos ecen WHO mo ali y da a o Sub-Saha-
an A ica only include high-quali y i al egis a ion da a o Mau i ius.
In he absence o i al egis a ion da a, age- and disease-speci ic mo ali y a es
o Sub-Saha an A ica ha e o be es ima ed. This equi es accu a e calcula ions o
o e all age-speci ic mo ali y a es o p o ide he “en elope” o o al dea hs ha
can hen be a ibu ed o di e en disease ca ego ies (Ma he s e al. 2004). When
da a on child mo ali y ( he likelihood o eaching age i e) and adul mo ali y ( he
likelihood o eaching age 60 gi en ha one has eached age 15) a e a ailable, modi-
ied logi li e ables can be used o es ima e mo ali y a es o all age g oups (Mu -
ay e al. 2003a, b a, b). Mu ay e al. (2003a, b) demons a e ha he p edic ion
me hod pe o ms well in simula ions i eliable da a on child mo ali y and adul
mo ali y o each sex a e a ailable; as he ela ionship be ween child and adul mo -
ali y di e s be ween coun ies and ac oss egions (see also Coale e al. 1983), hese
wo pieces o in o ma ion a e c ucial o gene a e he ull se o age‐ and sex-speci ic
mo ali y a es.
Because only child mo ali y da a was a ailable o he yea 2000 in Sub-Saha an
A ica, he co esponding le els o adul mo ali y we e essen ially assumed, using
his o ical model li e ables, as explained by Ma he s e al. (2004, p. 11): “Based on
he p edic ed le el o child mo ali y in 2002, he mos likely co esponding le el o
adul mo ali y (excluding HIV/AIDS dea hs whe e necessa y) was selec ed, along
wi h unce ain y anges, based on eg ession models o child e sus adul mo ali y
as obse ed in a se o almos 2000 li e ables judged o be o good quali y.”26 The e
is no independen way o con i ming hese adul mo ali y a es gene a ed by he li e
ables, le alone he ela ionship be ween male and emale adul mo ali y a es.
In mo e ecen yea s, he e has been some p og ess in he es ima ion o adul mo -
ali y a es (15–59). Since 2012, sex-speci ic adul mo ali y es ima es o he Global
Bu den o Disease s udies a e based on sibling su i al da a in coun ies whe e no
i al egis a ion da a a e a ailable, which in he case o Sub-Saha an A ica mos ly
come om Demog aphic and Heal h Su eys (Obe meye e al. 2010; Wang e al.
2020, see supplemen a y appendix 1, Sec . 2.3 and appendix Table1; Wo ld Heal h
O ganiza ion, 2020a,b). Albei no pe ec , hese adul mo ali y es ima es a e based
on ac ual da a om A ica and he e o e allow o a mo e eliable es ima ion o
25 See h ps:// www. who. in / da a/ gho/ da a/ hemes/ mo a li y- and- global- heal h- es im a es/ downl oad- he-
aw- da a- iles- o - he- who- mo a li y- da ab ase o “A ailabili y Con ains an Excel ile wi h he lis o
coun ies-yea s a ailable o he mo ali y and popula ion da a. Las upda ed: 15 Decembe 2019.”
26 No e ha he 2000 li e ables e e ed o he e and u he down in he pape by Mu ay e al. (2003a, b
a, b) a e om all o e he wo ld, no om A ica alone.
C.Ebe e al.
82 Page 22 o 26
age-speci ic mo ali y a es using modi ied logi model li e ables (Mu ay e al.
2003a,b; Wang e al. 2020, 2012, 2014; Wo ld Heal h O ganiza ion, 2020a,b).
The eno mous wo k o he Global Bu den o Disease p ojec in es ima ing age-
and disease-speci ic mo ali y a es o e he pas decades is o g ea alue and has
enabled he measu emen o he low o missing women in addi ion o nume ous
o he esea ch. S ill, he absence o i al egis a ion sys ems makes i impossible
o come up wi h p ecise age and disease-speci ic mo ali y a es. The a ailabili y
o i al egis a ion da a would immensely imp o e he measu emen o he age and
disease-speci ic low o missing women. When he low o missing women is es i-
ma ed using mo e ecen o , po en ially, in he u u e, using i al egis a ion da a, i
is impo an o unde s and ha his compa ison o e ime will be con ounded by he
changes in he measu emen o mo ali y. In con as , he s ock es ima es o missing
women ely on much mo e widely a ailable (and eliable) census in o ma ion and
can be used as a e e ence poin .
6 Conclusions
The low measu e o missing women calcula es excess emale mo ali y by age and
disease ha occu s in a ce ain ime pe iod, whe eas he p e ious li e a u e ocused
la gely on calcula ing he s ock measu e o missing women, i.e., he de ici among
ali e women a one poin in ime. Es ima ions o excess emale mo ali y by Ande -
son and Ray (2010) and Wo ld Bank (2012) sugges ha gende bias in mo ali y is
much la ge han p e iously ound (abou 4 o 5 million excess emale dea hs pe
yea s. a ound 100 million missing women in o al in 2000), is as se e e among
adul s as i is among child en in India, and is la ge in Sub-Saha an A ica han in
China and India. Howe e , hese es ima es depend on pa icula me hodological
choices wi h espec o he e e ence s anda d and he me hod o disease co ec ion.
I al e na i e, po en ially be e -sui ed, e e ence s anda ds a e used, he esul s align
wi h hose o he p e ious li e a u e on he s ock measu e o missing women. The
s ock measu e esul s show ha mo e women a e missing in China and India han in
Sub-Saha an A ica and ha , in China and India, missing women a e la gely a ib-
u able o excess emale mo ali y be o e bi h and/o he i s ew yea s o li e. In
addi ion o he ques ion o app op ia e e e ence s anda ds, a lack o i al egis a-
ion da a in mos de eloping coun ies makes i di icul o es ima e p ecise numbe s
o excess emale dea hs.
Gende bias in mo ali y is a se ious issue ha dese es u gen a en ion by poli-
cymake s. Acco ding o ou calcula ions, 1.5 o 1.8 million emales pe ish in excess
e e y yea . Al hough his numbe is much lowe han he o iginal low es ima es o
4 o 5 million yea ly excess emale dea hs, ou esul s sugges ha he ocus o he
li e a u e and associa ed policy p oposals on Sou h Asia and China, and on p e-bi h
and young child en, is jus i ied.
Ou es ima es con i m high excess emale mo ali y in Sub-Saha an A ica
among young adul s, mainly d i en by excess mo ali y om AIDS and ma e nal
mo ali y, which p o ides some in iguing u he de ail o he issue o gende bias
in mo ali y in Sub-Saha an A ica s udied so a (e.g.,Klasen 1996a, b; S edbe g
Coun ing missing women: a econcilia ion o low ands ock… Page 23 o 26 82
1996; Klasen and Wink 2002, 2003; Ainswo h and O e 1997). Fu he in es iga-
ion o gende bias in his age g oup in Sub-Saha an A ica should be a p io i y,
as should be con inuing e o s o imp o e i al da a om de eloping coun ies o
assess excess emale mo ali y mo e accu a ely in he u u e.
Acknowledgemen s We would like o hank Jö g Ankel-Pe e s, Je e Beh man, Sonia Bhalo a, Monica
Das Gup a, Angus Dea on, Ken Ha gen, Alan Lopez, Ashok Rai, Deb aj Ray, pa icipan s a semina s
and wo kshops a Ha a d, UPenn, Tu s, Dalhousie, and Gö ingen, and edi o Kompal Sinha and wo
anonymous e iewe s o help ul commen s and discussion. The au ho s decla e ha hey ha e no el-
e an o ma e ial inancial in e es s ha ela e o he esea ch desc ibed in his pape .
Funding Open Access unding enabled and o ganized by P ojek DEAL.
Da a A ailabili y The da a and ep oduc ion package o his a icle a e a ailable in he supplemen a y
ma e ials.
Open Access This a icle is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License,
which pe mi s use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any medium o o ma , as long
as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e
Commons licence, and indica e i changes we e made. The images o o he hi d pa y ma e ial in his
a icle a e included in he a icle’s C ea i e Commons licence, unless indica ed o he wise in a c edi line
o he ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons licence and you in ended
use is no pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will need o ob ain pe mis-
sion di ec ly om he copy igh holde . To iew a copy o his licence, isi h p://c ea i ecommons.o g/
licenses/by/4.0/.
Re e ences
Ab e aya J (2009) A e he e missing gi ls in he Uni ed S a es? E idence om bi h da a. Am Econ J
Appl Econ 1(2):1–34. h ps:// doi. o g/ 10. 1257/ app.1. 2.1
Adi janyee DR (1986) Suicide a emp s and suicides in India: c oss-cul u al aspec s. In J Soc Psychia y
32:64–73. h ps:// doi. o g/ 10. 1177/ 00207 64086 03200 208
Ainswo h M, O e AM (1997) Con on ing AIDS: public p io i ies in a global epidemic. Wo ld Bank
Policy Resea ch Repo , No. 17285, Washing on DC, The Wo ld Bank. h p:// docum en s. wo ld bank.
o g/ cu a ed/ en/ 21121 14687 79168 446. Accessed 13 Oc 2025
Akbulu -Yuksel M, Rosenblum D (2023) Es ima ing he e ec s o expanding ul asound use on sex selec-
ion in India. J De S ud. h ps:// doi. o g/ 10. 1080/ 00220 388. 2022. 21541 52
Ande son S, Ray D (2010) Missing women: age and disease. Re Econ S ud 77(4):1262–1300. h ps://
doi. o g/ 10. 1111/j. 1467- 937X. 2010. 00609.x
Ande son S, Ray D (2012) The age dis ibu ion o missing women in India. Econ Poli ical
Wkly47(47/48):87–95. h ps:// www. js o . o g/ s able/ 41720 413. Accessed 13 Oc 2025
Ande son S, Ray D (2019) Missing unma ied women. J Eu Econ Assoc 17(5):1585–1616
Anuk i i S (2018) Financial incen i es and he e ili y-sex a io ade-o . Am Econ J Appl Econ
10(2):27–57. h ps:// doi. o g/ 10. 1257/ app. 20150 234
As aw A, Klasen S, Lamanna F (2010) Gende gaps in pa en s’ inancing s a egy o hospi aliza ion o
hei child en: empi ical e idence om India. Heal h Econ 19(3):265–279. h ps:// doi. o g/ 10. 1002/
hec. 1468
Ba cellos SH, Ca alho LS, Lle as-Muney A (2014) Child gende and pa en al in es men s in India: a e
boys and gi ls ea ed di e en ly? Am Econ J Appl Econ 6(1):157–189. h ps:// doi. o g/ 10. 1257/
app.6. 1. 157
Bhalla K, Sangha i P (2020) Fi e- ela ed dea hs among women in India a e unde es ima ed. Lance
395(10226):779–780. h ps:// doi. o g/ 10. 1016/ S0140- 6736(20) 30060-X
C.Ebe e al.
82 Page 24 o 26
Bongaa s J, Guilmo o CZ (2015) How many mo e missing women? Excess emale mo ali y and p ena-
al sex selec ion, 1970–2050. Popul De Re 41(2):241–269. h ps:// doi. o g/ 10. 1111/j. 1728- 4457.
2015. 00046.x
Cen e s o Disease Con ol and P e en ion (2019) Es ima ed HIV incidence and p e alence in he
Uni ed S a es, 2010–2016. HIV Su eillance Supplemen al Repo 24 (No. 1). h p:// www. cdc. go /
hi / lib a y/ epo s/ hi - su e illan ce. h ml. Accessed 17 Ma 2021
Coale AJ (1991) Excess emale mo ali y and he balance o he sexes in he popula ion: an es ima e o
he numbe o “missing emales.” Popul De Re 17(3):517–523. h ps:// doi. o g/ 10. 2307/ 19719 53
Coale AJ, Banis e J (1994) Fi e decades o missing emales in China. Demog aphy 31(3):459–479.
h ps:// doi. o g/ 10. 2307/ 20617 52
Coale AJ, Demeny P, Vaughan B (1983) Regional model li e ables and s able popula ions, 2nd edn. Aca-
demic P ess, New Yo k. h ps:// doi. o g/ 10. 1016/ C2013-0- 07295-7
C immins EM, P es on SH, Cohen B (2011) Explaining di e gen le els o longe i y in high-income
coun ies. Washing on, D.C., The Na ional Academies P ess. h ps:// doi. o g/ 10. 17226/ 13089
Das Gup a M (1987) Selec i e disc imina ion agains emales in u al Punjab. India Popula ion and
De elopmen Re iew 13(1):77–100. h ps:// doi. o g/ 10. 2307/ 19721 21
Das Gup a M (2005) Explaining Asia’s “missing women”: a new look a he da a. Popul De Re
31(3):529–535. h ps:// doi. o g/ 10. 1111/j. 1728- 4457. 2005. 00082.x
Das Gup a M (2006) Cul u al e sus biological ac o s in explaining asia’s “missing women”: esponse o
os e . Popul De Re 32(2):328–332. h ps:// doi. o g/ 10. 1111/j. 1728- 4457. 2006. 00121.x
Das Gup a M, Chung W, Shuzhou L (2009) Is he e an incipien u na ound in Asia’s “missing gi ls”
phenomenon?. Policy Resea ch Wo king Pape Se ies 4846. The Wo ld Bank. h ps:// hdl. handle. ne /
10986/ 4040. Accessed 13 Oc 2025
Dea on A (2004) Heal h in an age o globaliza ion. In: Collins S, G aham C (eds) B ookings T ade
Fo um. Washing on, D.C., The B ookings Ins i u ion. h ps:// doi. o g/ 10. 3386/ w10669
Dyson T, Moo e M (1983) On kinship s uc u e, emale au onomy, and demog aphic beha io in India.
Popul De Re 9(1):35–60. h ps:// doi. o g/ 10. 2307/ 19728 94
Ebe C, Vollme S (2022) Gi ls unwan ed - he ole o pa en s’ child-speci ic sex p e e ence o chil-
d en’s ea ly men al de elopmen . J Heal h Econ 82:102590. h ps:// doi. o g/ 10. 1016/j. jheal eco. 2022.
102590
Hi oeh U, Appleby L, Mo ensen PB, Dunn G (2001) Dea h by homicide, suicide, and o he unna u-
al causes in people wi h men al illness: a popula ion-based s udy. Lance 358(9299):2110–2112.
h ps:// doi. o g/ 10. 1016/ S0140- 6736(01) 07216-6
Hogan MC, Fo eman K J, Nagha i M, Ahn S Y, Wang M, Makela SM, Lopez AD, Lozano R, Mu -
ay CJL (2010) Ma e nal mo ali y o 181 coun ies, 1980–2008: a sys ema ic analysis o p og ess
owa ds Millennium De elopmen Goal 5. Lance 375(9726):1609–1623. h ps:// doi. o g/ 10. 1016/
s0140- 6736(10) 60518-1
Ili e J (2006) The A ican AIDS epidemic: a his o y. Ohio Uni e si y P ess, A hens. h ps:// doi. o g/ 10.
14296/ RiH/ 2014/ 2038
Ja adeka S, Saxena K (2025) The seen and unseen: he unin ended impac o a condi ional cash ans e
p og am on p ena al sex selec ion. J Popul Econ 38:32. h ps:// doi. o g/ 10. 1007/ s00148- 025- 01091-6
Jayachand an S, Kuziemko I (2011) Why do mo he s b eas eed gi ls less han boys? E idence and impli-
ca ions o child heal h in India. Q J Econ 126(3):1485–1538. h ps:// doi. o g/ 10. 1093/ qje/ qj 029
Jayachand an S, Rohini P (2017) Why a e Indian child en so sho ? The ole o bi h o de and son p e e -
ence. Am Econ Re 107(9):2600–2629. h ps:// doi. o g/ 10. 1257/ ae . 20151 282
Johansson S (1984) De e ed in an icide: excess emale mo ali y in childhood. In: Haus a e G, H dy S
(eds) In an icide. Aladin, New Yo k, pp 463–486
Johansson S (1991) Wel a e, mo ali y and gende . Con inui y and change in he explana ion o male/
emale mo ali y di e en ials o e h ee cen u ies. Con in Chang 6(2):135–177. h ps:// doi. o g/ 10.
1017/ S0268 41600 00013 14
Jones TR, Milling on MJ, P ice J (2023) Changes in pa en al gende p e e ence in he USA: e idence
om 1850 o 2019. J Popul Econ 36(4):3057–3070. h ps:// doi. o g/ 10. 1007/ s00148- 023- 00957-x
Kalben BB (2000) Why men die younge . N Am Ac ua J 4(4):83–111. h ps:// doi. o g/ 10. 1080/ 10920
277. 2000. 10595 939
Klasen S (1994) Missing women econside ed. Wo ld De 22(7):1061–1071. h ps:// doi. o g/ 10. 1016/
0305- 750X(94) 90148-1
Klasen S (1996) Nu i ion, heal h, and mo ali y in sub Saha an A ica: is he e a gende bias? J De S ud
32(6):913–932. h ps:// doi. o g/ 10. 1080/ 00220 38960 84224 46
